You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Apache Spark (Jira)" <ji...@apache.org> on 2021/10/11 15:45:00 UTC
[jira] [Assigned] (SPARK-36867) Misleading Error Message with
Invalid Column and Group By
[ https://issues.apache.org/jira/browse/SPARK-36867?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Apache Spark reassigned SPARK-36867:
------------------------------------
Assignee: Apache Spark
> Misleading Error Message with Invalid Column and Group By
> ---------------------------------------------------------
>
> Key: SPARK-36867
> URL: https://issues.apache.org/jira/browse/SPARK-36867
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 3.1.2
> Reporter: Alan Jackoway
> Assignee: Apache Spark
> Priority: Major
>
> When you run a query with an invalid column that also does a group by on a constructed column, the error message you get back references a missing column for the group by rather than the invalid column.
> You can reproduce this in pyspark in 3.1.2 with the following code:
> {code:python}
> from pyspark.sql import SparkSession
> spark = SparkSession.builder.appName("Group By Issue").getOrCreate()
> data = spark.createDataFrame(
> [("2021-09-15", 1), ("2021-09-16", 2), ("2021-09-17", 10), ("2021-09-18", 25), ("2021-09-19", 500), ("2021-09-20", 50), ("2021-09-21", 100)],
> schema=["d", "v"]
> )
> data.createOrReplaceTempView("data")
> # This is valid
> spark.sql("select sum(v) as value, date(date_trunc('week', d)) as week from data group by week").show()
> # This is invalid because val is the wrong variable
> spark.sql("select sum(val) as value, date(date_trunc('week', d)) as week from data group by week").show()
> {code}
> The error message for the second spark.sql line is
> {quote}
> pyspark.sql.utils.AnalysisException: cannot resolve '`week`' given input columns: [data.d, data.v]; line 1 pos 81;
> 'Aggregate ['week], ['sum('val) AS value#21, cast(date_trunc(week, cast(d#0 as timestamp), Some(America/New_York)) as date) AS week#22]
> +- SubqueryAlias data
> +- LogicalRDD [d#0, v#1L], false
> {quote}
> but the actual problem is that I used the wrong variable name in a different part of the query. Nothing is wrong with {{week}} in this case.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org